The MFEA Encoder

For our robotic arm to work effectively, a rotary encoder had to be found. A baseline specification for such an encoder was established by compiling a system error budget of which the rotary encoder forms part. From this, it was determined that an angular encoder was required with the following baseline specification:

1. 16 bit resolution.
2. 0.112 mrad accuracy.
3. 0.056 mrad repeatability.
4. Ability to work in harsh environments without performance degradation.​

Problem was we could not find an encoder satisfying such a baseline. As a result, we decided to build our own. And after many months of development we were able to produce an encoder able to surpass the performances of current best in class encoders available on the market. During a series of tests at room temperature the static accuracy of our prototype ADM3 was measured and found to be in the order of 0,07 mrad with repeatability of 0.035 mrad.

We were ecstatic with our prototype encoders performance. So naturally we decided to patent the technology. In 2021 we patented our encoder technology (Patent Pending 2021/08462). We called our encoder MFEA; short for Magnetic Field Effect Angular Encoder leaving out the “E” of the word “Encoder” for practical purposes.

We discovered a new way to measure angle accurately and reliably. And do so in harsh environments where oil, grease, dust and moisture are prevalent; incidentally the same environment typically experienced in factories and industrial automation. And the electronic components and assembly processes of our encoder are not exotic, but entails straight forward PCB electronics applied in a clever way.

How does it work?

Here follows a basic explanation the MFEA encoder concept:

As the spindle rotates, the “overlap area” formed between the inductor coils (in red) and the wobbling ferrite toroid (in blue) varies. This variation in overlap area results in a variation of inductance resulting in the frequency variation of two separate colpitts oscillators.

Each oscillator is buffered with a voltage comparator, producing two high-speed digital signals that can be transmitted with simple wiring over several meter distance to two 16-bit binary counters in a common micro-processor.

The counters determine the number of pulses produced by each oscillator during a precise period of time. The microprocessor uses these count values to determine with high precision the angle of the spindle.

The Auto-Align MFEA Encoder

Current angular encoders, including best in class offer very little tolerance to radial and axial misalignment. As a result the robotics and automation industry is plagued with difficult installation procedures requiring utmost care and precision when installing such a device. And in cases where installation tolerance cannot be guaranteed, radial and axial misalignment needs to be accommodated by re-calibrating the sensor in-situ. This is a lengthy, cumbersome process; often leading to significant downtime. And in some cases (such as in space and high pressure under-water environments) calibration becomes practically impossible. Our Angular Encoder ADM3 is no exception to the rule. Although ADM3 outperforms best in class encoders, it is also plagued by the same installation and calibration problems, vividly demonstrated while installing it in our Forearm Actuator ADM3 model.

As a result, our design team considered the possibility of developing an encoder able to tolerate substantial radial and axial misalignment so as to negate the need to calibrate. Such an encoder would be a game-changer; and would most certainly be high in demand. Landman Robotics was at a cross-road: should the Company focus efforts in further developing the Angular Encoder ADM3 model and produce the first Engineering Development Model (EDM) or should they focus their efforts in developing a way of addressing the installation problem?

Luckily a breakthrough was made. After many months of hard work, the nut was finally cracked. This gave impetus to the development of our Angular Encoder ADM5, the encoder consequently called the “Auto-Align” encoder.

ADM5 is based on the same principles of operation as used in ADM3, but includes an algorithm that automatically determines the misalignment between rotor and stator, thus affecting the auto-align function as described above. A Field Programmable Gate Array (FPGA) containing a CAN bus interface and dedicated circuitry for running the auto-align algorithm at high speed was introduced in FPGA fabric. The auto-align algorithm has been proven to work in Simulink simulations. These simulations showed that the algorithm can lock onto a solution within 20 microseconds proving the idea to be viable. Angular Encoder ADM5 model will be able to accommodate sizeable radial and axial misalignments in the order of ±1° and ±2mm respectively. And the device would actively calibrate (or auto-align for a better choice of words) itself, negating the need to ever calibrate the sensor in the traditional sense.

ADM5 is currently being developed and nearing completion. We are planning to have working samples completed by the end of 2022. These will be send to the MFEA Production Houses for evaluation.